*Retribution versus Rehabilitation for Children within Insurgency: Public Attitudes toward ISIS-affiliated Youth in Mosul, Iraq

*Replication Instructions

*Vera Mironova and Sam Whitt

*Below are instructions for replicating all manuscript and online appendix tables and figures in STATA using the WIDE format of the dataset "TPV Children within insurgency replication data wide format.dta". Please contact Sam Whitt (swhitt@highpoint.edu) for questions regarding data replication.Use the appropriate "wide format" do file to replicated different components of the manuscript.

*Note: You may need to install STATA packages for cibar, catcibar  and iebaltab commands. Use findit with command name to identify and download the appropriate packets to install. To install the catcibar command enter the following:
*ssc install cibar
*net install catcibar, from("https://aarondwolf.github.io/catcibar")

*Note: In addition, some graphs require additional formatting using filename.grec files with the graph play command. To format a graph, simply run the command to generate the graph in the do file in STATA, then open the "Graph Editor" in STATA and click on the GREEN "Play Recording" button, then select "Browse" to select the grec file from among Replication files. The name of the grec file is indicated in the note below the graph command in the do file for the specific graph you wish to format. This should automatically format the graph, which you may then save to a location of your choosing.


*Manuscript Text

*Use wide format replication data for all text replication

*"In total, 358 subjects completed our study between April 22 and May 6, 2019."

tab date

*"Respondents ranged in age from 18-61, the vast majority of which were male (93%)."
sum age
tab female

*"Figure 1 indicates responses to these inquiries. First, it shows that the overwhelming majority of Moslawis (97.8%) identify eighteen as the threshold age of adulthood."

tab adultage

*"The most common explanation for this age is correctly based on Iraqi law (90%), but some respondents also provided extra-legal justifications for adulthood based on religion (23%), agency (14%) and autonomy (17%), development (4%), and their own experiences with minors (25%)."

sum reasonslaw-reasonsexperience

*Index of Extra-legal determinants of Adulthood
alpha reasonsreligion-reasonsexperience

*"Figure 1 indicates that many respondents are willing to punish children under the age of eighteen as adults for ISIS involvement, in violation of Iraqi law. Only 30% state that the age at which ISIS participants should be punished as adults must be equivalent to their own prior reported threshold of adulthood. The majority (68%) would lower the age of punishment to one or more years below their adult threshold when punishing ISIS-affiliated youth."

tab ageofpunishment if adultage==18
tab agedifference

*"Two sample t-tests indicate that parental coercion reduces punishment severity (t = 5.97, p<0.0000), and the effect size is considerably large (Cohen's d=0.69)."

ttest ahmadfnf, by (ahmadtxt) unequal
cohend ahmadfnf ahmadtxt

*"We find that respondents are divided on this question with 58% agreeing or strongly agreeing that putting children of ISIS-affiliated families in a state-run facility would reduce radicalization and 42% disagreeing or strongly disagreeing."

tab withstate
tab revwithstate

*"To evaluate H3, we measure threat perceptions by asking respondents whether "Children in ISIS could be just as dangerous as adults in the future" [1 = strongly disagree to 4 = strongly agree]. The vast majority somewhat agree (37.8%) or strongly agree (59.7%) that this was a concern."

tab revchildthreat

*"This item links rehabilitation prospects with concerns about parental socialization effects driving radicalization. Respondents are again divided with 54.1% agreeing that this could prevent future radicalization and 45.9% disagreeing."

tab withparents
tab revwithparents


*"This item examines support for more rehabilitation efforts to prevent radicalization. The vast majority either strongly agree (53.8%) or somewhat agree (43.1%) that this is needed to reduce future radicalization."

tab revreeducation

*"Figure 4a-d. report the frequency distributions of responses to each item by treatment. Mann-Whitney tests indicate significant differences in response distribution across treatments below the p<0.05 level with moderate effect sizes ranging from Cohen's d=0.53, 0.48, 0.76, and 0.23 respectively."

ranksum revsamidanger, by( samitxt)
ranksum revsamiradical, by( samitxt)
ranksum revsamirevenge, by( samitxt)
ranksum revsamionlychild, by( samitxt)

cohend revsamidanger samitxt
cohend revsamiradical samitxt
cohend revsamirevenge samitxt
cohend revsamionlychild samitxt

*"In Figure a, b, and c, respondents indicate that the child in the local treatment is a greater danger, more likely to become radical, and more likely to seek revenge, than in the foreign treatment. We combine these items into a common index of threat perception (see online appendix for details)."

*Child Threat Perception Index Variable

factor revsamidanger-revsamiradical
alpha revsamidanger-revsamiradical


*"Figure 4d shows that many respondents strongly agree (56.6%) or somewhat agree (41.5%) with the statement that Sami is only a child and should not be punished for his father's crimes, even if he may be also regarded as a future threat in Figure 4a-c."

tab revsamionlychild


*Manuscript Tables and Figures


*Figure 1. Determinants of adulthood

*Figure 1a. Age at which children should be punished as adults.

twoway  (histogram ageofpunishment, discrete percent) (histogram adultage, width(1) discrete percent fcolor(none) lcolor(black)), legend(position(6))

graph save g1, replace

*Figure 1b. Determinants of adulthood

graph bar reasonslaw reasonsexperience reasonsreligion reasonsindependence reasonsright reasonsdevelopment , blabel(bar, position(outside) format(%5.2g)) legend(position(6))

graph save g2, replace
graph combine "g1.gph" "g2.gph"

*Note additional formatting requires the "Figure 1 wide formatting.grec" file with the command graph play "Figure 1 wide formatting.grec"

*Figure 2. Mean punishment preferences for adults and children
* see long format dataset

*Figure 3. Punishment preferences with parental coercion treatments.

*Figure 3a.
cibar ahmadfnf, over1(ahmadtxt)
graph save g3, replace

*Figure 3b. 
twoway  (histogram ahmadfnf if ahmadtxt==1, discrete percent) (histogram ahmadfnf if ahmadtxt==2,discrete percent fcolor(none) lcolor(black))
graph save g4, replace
graph combine "g3.gph" "g4.gph"

*Note additional formatting requires the "Figure 3 wide formatting.grec" file with the command graph play "Figure 3 wide formatting.grec"

*Table 1. Punishment preferences for adults and children (OLS regression).

*Model 1 (see Long format)

*Model 2 (see long format)

*Model 3 
reg childfought revchildsupport revnopunchildfighter extralegal ageofpunishment female age education professional laborer unemployed income i.religion revimpreligion i.ethnicity revimpethnicity, robust

*Figure 4. Threat perception and rehabilitation prospects of a young child

*Figure 4a. 
twoway  (histogram revsamidanger if samitxt==1, discrete percent) (histogram revsamidanger if samitxt==2,discrete percent fcolor(none) lcolor(black))
graph save g5, replace

*Figure 4b.
twoway  (histogram revsamiradical if samitxt==1, discrete percent) (histogram revsamiradical if samitxt==2,discrete percent fcolor(none) lcolor(black))
graph save g6, replace

*Figure 4c.
twoway  (histogram revsamirevenge if samitxt==1, discrete percent) (histogram revsamirevenge if samitxt==2,discrete percent fcolor(none) lcolor(black))
graph save g7, replace

*Figure 4d.
twoway  (histogram revsamionlychild if samitxt==1, discrete percent) (histogram revsamionlychild if samitxt==2,discrete percent fcolor(none) lcolor(black))
graph save g8, replace

graph combine "g5.gph" "g6.gph" "g7.gph" "g8.gph"

*Note additional formatting requires the "Figure 4 wide formatting.grec" file with the command graph play "Figure 4 wide formatting.grec"

*Table 2. Opposition to reintegration into society (OLS regression).

reg revwithstate revchildthreat revwithparents revreeducation   female age education professional laborer unemployed income i.religion revimpreligion i.ethnicity revimpethnicity, robust

reg revwithstate c.samithreat#samitxt  female age education professional laborer unemployed income i.religion revimpreligion i.ethnicity revimpethnicity, robust

reg revwithstate c.revsamionlychild#samitxt female age education professional laborer unemployed income i.religion revimpreligion i.ethnicity revimpethnicity, robust


*Appendix Replication

*Mosul Sampling Locations

tab locale

*Summary of Variables in Manuscript Analysis

sum  childfought childworked adultfought adultworked ageofpunishment reasonslaw-reasonsexperience extralegal revchildsupport revnopunchildfighter revwithstate revchildthreat revwithparents revreeducation

sum ahmadfnf if ahmadtxt==1
sum ahmadfnf if ahmadtxt==2

sum c.samithreat#samitxt if samitxt==1
sum c.samithreat#samitxt if samitxt==2

sum c.revsamionlychild#samitxt  if samitxt==1
sum c.revsamionlychild#samitxt  if samitxt==2

*Figure 1 General Punishment Preferences for ISIS Fighters (Child/Adult)

*see long format

*Figure 2 General Punishment Preferences for ISIS Workers (Child/Adult)

*see long format

*Figure 3. Ahmad 10-year old fighter vignette

twoway  (histogram ahmadfnf if ahmadtxt==1, discrete percent) (histogram ahmadfnf if ahmadtxt==2,discrete percent fcolor(none) lcolor(black))

*Note additional formatting required

*Experimental Balance Tests across Treatment Groups

ksmirnov female, by(surveyversion12)
ksmirnov age, by(surveyversion12)
ksmirnov education, by(surveyversion12)
ksmirnov professional, by(surveyversion12)
ksmirnov laborer, by(surveyversion12)
ksmirnov unemployed, by(surveyversion12)
ksmirnov income, by(surveyversion12)
ksmirnov religion, by(surveyversion12)
ksmirnov ethnicity, by(surveyversion12)

*Treatment Effect Estimation for Child Fighter with/wo Father Survey Vignette (Ahmad)

teffects ra (ahmadfnf income) (ahmadtxt)
teffects ipwra (ahmadfnf) (ahmadtxt income)
teffects psmatch (ahmadfnf) (ahmadtxt income)

*Treatment Effect Estimation for Child of Local/Foreign Fighter Survey Vignette (Sami, Threat Perception Index)

teffects ra (samithreat income) (samitxt)
teffects ipwra (samithreat) (samitxt income)
teffects psmatch (samithreat) (samitxt income)

*Manuscript Table 1. Robustness Check (Ordered Probit Regression)

*Model 1 (see Long format)

*Model 2 (see long format)

*Model 3 (wide format)
oprobit childfought revchildsupport revnopunchildfighter extralegal ageofpunishment female age education professional laborer unemployed income i.religion revimpreligion i.ethnicity revimpethnicity, robust

*Manuscript Table 2. Robustness Check (Ordered Probit Regression)

oprobit revwithstate revchildthreat revwithparents revreeducation   female age education professional laborer unemployed income i.religion revimpreligion i.ethnicity revimpethnicity, robust

oprobit revwithstate c.samithreat#samitxt c.revsamionlychild#samitxt  female age education professional laborer unemployed income i.religion revimpreligion i.ethnicity revimpethnicity, robust

*Robustness Checks for Locational Fixed Effects

*Punishment Preferences (OLS, Locational Effects)

*See Long format

*Opposition to Reintegration into Society (OLS, Locational Effects)

reg revwithstate revchildthreat revwithparents revreeducation female age education professional laborer unemployed income i.religion revimpreligion i.ethnicity revimpethnicity westmosul, robust

reg revwithstate c.samithreat#samitxt c.revsamionlychild#samitxt  female age education professional laborer unemployed income i.religion revimpreligion i.ethnicity revimpethnicity westmosul, robust

log close